A Hierarchical Task Allocation Method Based on Task Case and its Application to Multi-Robot Hunting

Abstract

This paper proposes a hierarchical task allocation method, based on the matching of task cases. On the basis of hierarchical organization form of multi-robot system, the tasks are allocated by case matching, which is based on the combination of contract net and acquaintance net. After a new task is arrived, its features are synthesized and then it is matched from task cases library. When the stored items of matched case in the library are enough, high-layer manager will allocate the task to the selected low-layer manager directly, or else, it adopts contract net based allocation to each low-layer manager, who will bid to high-layer manager directly based on acquaintance net. The execution result of selected robots will be returned to the high-layer manager, who will update the cases library. The method is applied to multi-robot hunting to show its validity.